Intravenous Amisulpride Does Not Meaningfully Prolong the QTc Interval at Doses Effective for the Management of Postoperative Nausea and Vomiting : Anesthesia & Analgesia

Secondary Logo

Journal Logo

Original Research Articles: Original Clinical Research Report

Intravenous Amisulpride Does Not Meaningfully Prolong the QTc Interval at Doses Effective for the Management of Postoperative Nausea and Vomiting

Fox, Gabriel M. MB BChir*; Albayaty, Muna MBChB, FFPM, MSc; Walker, Joanna L. BSc*; Xue, Hongqi PhD; Darpo, Borje MD, PhD

Author Information
Anesthesia & Analgesia 132(1):p 150-159, January 2021. | DOI: 10.1213/ANE.0000000000004538
  • Free



  • Question: Does a 10-mg dose of intravenous amisulpride, given alone or with ondansetron, prolong the QT interval to a clinically relevant extent?
  • Findings: The effect on the heart rate-corrected QT (QTc) interval of a single 10-mg dose of intravenous amisulpride alone and in combination with 4-mg ondansetron was below the 10-millisecond threshold of concern identified by regulatory agencies.
  • Meaning: Intravenous amisulpride is unlikely to contribute a meaningful risk of serious cardiac arrhythmias, such as torsade de pointes, when given at doses effective for the management of postoperative nausea and vomiting (PONV).

Managing postoperative nausea and vomiting (PONV) is an important priority for the surgical team. Patients at moderate-to-high risk for PONV should receive ≥1 prophylactic antiemetics of different classes before or during surgery; where these fail to prevent PONV, treatment options are limited.1 Previously, droperidol, a dopamine D2-antagonist, was widely used for PONV management. However, since it received a boxed warning from Food and Drug Administration (FDA) in 2001 due to events of torsade de pointes (TdP) arising from QT interval prolongation, its use has declined markedly, especially in the United States.2 Other dopamine antagonists used as antiemetics, such as haloperidol, also prolong the QT interval. An effective dopaminergic antiemetic with no clinically significant QT effect would be a valuable addition to the PONV armamentarium.

Amisulpride is an equipotent D2 and D3 antagonist shown to be effective at low intravenous (IV) doses in both the prophylaxis and treatment settings. A 5-mg dose was effective at preventing PONV, either alone or when given in combination with another antiemetic with a different mechanism of action.3–5 A 10-mg dose was effective for treating established PONV, whether or not previous prophylaxis had been given with an antiemetic of a different class.6,7

In a “thorough QT” (TQT) study—a placebo- and positive-controlled study designed to exclude a small QT interval corrected for heart rate (QTc) effect8,9—IV amisulpride was shown to have a dose- and concentration-dependent effect on the QT interval.10 At 5 mg, the largest mean effect on the placebo-corrected change-from-baseline QT interval corrected for heart rate (HR) using Fridericia’s formula (QTcF) (ΔΔQTcF) was 5.0 milliseconds (90% confidence interval [CI], 2.8–7.1 milliseconds), hence below the 10 milliseconds threshold of regulatory concern. A supratherapeutic dose of 40 mg resulted in a largest mean ΔΔQTcF of 23.4 milliseconds (90% CI, 21.3–25.5 milliseconds).

Subsequently, it was shown that 10-mg IV amisulpride was effective for treating established PONV. To characterize this clinically important dose fully, we conducted this randomized, double-blind, crossover study in healthy, adult volunteers, with the primary objective of quantifying its QT effect, both as a single agent and when combined with 4-mg IV ondansetron, an antiemetic known to cause QT prolongation, which many patients requiring rescue after failed prophylaxis will have received. The hypothesis tested was that an effect on the QTc interval exceeding 10 milliseconds could be excluded at the average plasma concentration delivered by a single 10-mg IV dose of amisulpride, alone and when coadministered with 4-mg IV ondansetron. As a secondary objective, the safety of a repeat 10-mg dose was evaluated; pharmacokinetic data used in the concentration-QTc (C-QTc) analysis are also provided.


This double-blind, randomized, placebo-controlled, crossover study was conducted at a single phase 1 unit in London, United Kingdom, in July–August 2018. Before subject enrolment, the study was registered on the EudraCT database (January 12, 2018; number: 2018-000154-21) and on (July 10, 2018; number: NCT03583489; Principal Investigator: M.A.). An independent ethics committee approved the study. Written informed consent was obtained from all subjects before enrolment. This manuscript adheres to the applicable Consolidated Standards of Reporting Trials (CONSORT) guidelines.

Healthy male and female subjects were eligible for inclusion if they were 18–65 years of age, had a body mass index of 18–30 kg/m2, and agreed to use adequate contraception. Typical exclusion criteria for studies in healthy subjects were applied and included electrocardiogram (EKG) abnormalities (eg, QTc >450 milliseconds), history of clinically significant cardiovascular disease, syncope, electrolyte disturbances, family history of congenital long QT syndrome, and any clinically important abnormalities in rhythm, conduction, or morphology of resting EKG that could have interfered with the interpretation of QTc interval changes.

Subjects were screened for eligibility within 28 days before dosing. Eligible subjects were admitted to the study unit the afternoon before the first day of dosing and remained resident in the unit until after the final blood draw of the last dosing period. Subjects went through 3 dosing periods in random order, according to a randomization code generated by biostatisticians at the study site, with each period comprising 2 simultaneous 1-minute infusions, followed by a further 1-minute infusion 2 hours later, all into a peripheral vein. The study periods were as follows: (1) amisulpride 10 mg and placebo, followed by amisulpride 10 mg 2 hours later; (2) amisulpride 10 mg and ondansetron 4 mg, followed by placebo 2 hours later; and (3) 2 simultaneous placebo infusions, followed by placebo 2 hours later. The placebo comprised the same excipients as IV amisulpride. Subjects fasted for at least 8 hours before the first dosing of each period. There was a washout period of at least 48 hours between each period. All study drugs were prepared by an unblinded pharmacy technician to look identical, so that neither study site staff nor subjects could distinguish what was being administered.

In each period, venous blood was drawn for measurement of study drug concentrations before the start of dosing and at 1 (end of infusion), 5, 10, 20, 30, 60, 120 (immediately before second infusion), 121 (end of second infusion), 125, 130, 140, 150, 180, 240, and 360 minutes after the start of study drug dosing. Plasma samples for determination of amisulpride concentration were analyzed by LGC Ltd (Fordham, UK), using a validated liquid chromatography/tandem mass spectrometry/mass spectrometry (LC/MS/MS) method with liquid–liquid extraction, validated over the calibration range 2–2000 ng/mL with a lower limit of quantification of 2 ng/mL. The intra- and interassay coefficients of variation at 6, 60, and 1500 ng/mL, in the presence or absence of 250 ng/mL ondansetron, were all ≤7.4%. Quantitation of ondansetron in plasma was done by PPD Laboratories (Middleton, WI) using high-performance liquid chromatography (HPLC) with MS/MS detection, validated between 0.5 and 250 ng/mL. PPD Laboratories also performed a check to confirm an absence of interference between ondansetron and amisulpride in the assay.

Continuous 12-lead EKGs were recorded using a Global Instrumentation M12R digital recorder (Manlius, NY), and EKG extractions were made in each period before dosing at 3 time-points and postdose at the same time as for drug concentration assay. Extractions were done using the “TQT Plus method” (ERT, Rochester, NY),11 a computer-assisted and statistical process enabling extraction of EKGs with the lowest HR variability and noise within the protocol-specified extraction time window. At each protocol-specified time-point, 10 EKG replicates were extracted from a 5-minute window, with the subject in a supine or semirecumbent, quiet position. Expert precision QT analysis was performed on all analyzable (nonartifact) beats in the 10 EKG replicates.11 Statistical quality control procedures were used to review and assess all beats and identify “high” and “low” confidence beats using several criteria, including QT, QTc, or RR values above or below certain (biologically unlikely) thresholds and rapid changes in QT, QTc, or RR from beat to beat. Measurements of all primary EKG parameters (QT, QTc, and RR) in all recorded beats of all replicates deemed “high confidence” were performed using COMPAS software (eResearch Technologies Inc, Rochester, NY). All low confidence beats were reviewed manually and adjudicated using pass–fail criteria. The final QC assessment was performed by a cardiologist. The beats found acceptable by manual review were included in the analysis. The median QT, QTc (corrected for HR using Fridericia’s formula: ), and RR value from each extracted replicate was calculated, and then the mean of all available medians from a nominal time-point was used as the subject’s reportable value at that time-point. Categorical T-wave morphology analysis and the measurement of PR and QRS intervals were performed with a semiautomated technique in 3 of the 10 EKG replicates at each time-point.

Blood and urine samples were collected for hematology and biochemistry analysis and urinalysis at 6 hours after the start of dosing in each period. Adverse events were recorded from clinic admission until discharge from the study and judged as mild, moderate, severe, or life threatening according to standard criteria (see Table 2 footnote).

Statistical Methods

The primary end point of the study was ΔΔQTcF interval. Baseline was calculated as the average from the 3 EKG time-points before dosing (−45, −30, and −15 minutes) from the continuous Holter recording. The primary analysis was a C-QTc analysis based on change-from-baseline QTcF (ΔQTcF). The relationship between plasma concentrations of amisulpride and ondansetron and ΔQTcF was investigated by a linear mixed-effects modeling approach, with a full model that included ΔQTcF as the dependent variable, time-matched plasma concentrations of amisulpride and ondansetron and their interaction term (without quadratic terms) as the explanatory variates, centered baseline QTcF (ie, baseline QTcF for individual subject subtracting the population mean baseline QTcF for all subjects) as an additional covariate, treatment (0 for placebo and 1 for active treatment periods) and time (ie, time-point) as categorical factors, and subject as random effect for both intercept and slopes.12 One reduced model was also derived which involved amisulpride alone as an explanatory variate.

A ‘by-time-point’ analysis for QTcF was performed as a secondary analysis, using a linear mixed-effects model with change-from-baseline QTcF (ΔQTcF) as the dependent variable, period, sequence, time (ie, time-point: categorical), treatment (amisulpride alone, amisulpride with ondansetron, and placebo), and time-by-treatment interaction as fixed effects and baseline QTcF as a covariate. An unstructured covariance matrix was specified for the repeated measures at postdose time-points for subject within treatment period. From this analysis, the least-squares mean and 2-sided 90% CI were calculated for the contrast “amisulpride with/without ondansetron versus placebo” for amisulpride alone or amisulpride with ondansetron, at each postdose time-point, separately. For HR, PR, and QRS interval, the analysis was based on the change-from-baseline postdosing (change-from-baseline HR [ΔHR], change-from-baseline PR [ΔPR], and change-from-baseline QRS [ΔQRS]). The same (“by-time-point” analysis) model was used as described for QTcF. In both the C-QTc and the by-time-point analyses, 90% CIs were used based on available clinical guidance on the evaluation of drug effects on EKG parameters.12,13

Hysteresis was investigated using joint graphical displays of the least-squares mean difference between ΔQTcF under amisulpride alone (or amisulpride with ondansetron) and under placebo (ΔΔQTcF) for each postdose time-point and the mean concentrations of amisulpride (and ondansetron) at the same time-points. In addition, hysteresis plots were given for least-squares mean ΔΔQTcF and either of the mean concentrations.

Categorical analysis of outliers included increases in absolute QTcF interval values >450 and ≤480 milliseconds, >480 and ≤500 milliseconds, and >500 milliseconds and changes from predose baseline of >30 and ≤60 milliseconds, and >60 milliseconds; increase in PR interval from predose baseline >25% to a PR >200 milliseconds; increase in QRS interval from predose baseline >25% to a QRS >120 milliseconds; decrease in HR from predose baseline >25% to HR <50 beats per minute (bpm); and increase in HR from predose baseline >25% to HR >100 bpm. Categorical analysis of T-wave morphology and U-wave presence focused on change from baseline.

The pharmacokinetic parameters maximum observed plasma concentration (Cmax) and area under the concentration versus time curve from time zero extrapolated to infinity (AUC[0–∞], representing total drug exposure over time) were calculated from raw concentration data using the R package “PKNCA.”14

Based on substantial recent experience from studies using C-QTc analysis, a sample size of 24 provides 80% power to exclude that amisulpride with ondansetron causes more than a 20-millisecond QTc effect at clinically relevant plasma levels, as shown by the upper bound of the 2-sided 90% CI of the model-predicted QTc effect (ΔΔQTcF) at the observed geometric mean Cmax in the study. The calculation assumes an underlying effect of amisulpride with ondansetron of 15 milliseconds, estimated from previous studies on the QT effect of the drugs,10,15,16 and a standard deviation (SD) of ΔQTcF of 7 milliseconds using the described technique for measurement of the QT interval.11 In addition, the sample size can be estimated approximately using a simple paired t test for equivalence. Under the assumption that the QTc effect is 15 milliseconds for amisulpride with ondansetron and 0 milliseconds for placebo with a SD of ΔQTcF of 7 milliseconds for each treatment and that “No Effect” is to be concluded if the upper bound of the 2-sided 90% CI of ΔΔQTcF is lower than 20 milliseconds, 24 subjects would provide 80% power with a 1-sided α of 5% in a paired t test.


A total of 30 subjects were enrolled and randomized, and 29 completed all 3 dosing periods. One subject completed the first period (amisulpride plus ondansetron) but withdrew during the second (placebo) due to infusion site pain during administration.

Baseline values of EKG parameters were very similar for each dosing period, with mean HR between 59.8 and 61.2 bpm, mean QTcF 405.0–406.6 milliseconds, mean PR 157.7–157.8 milliseconds, and mean QRS 105.3–105.8 milliseconds.

Amisulpride plasma concentration profiles (arithmetic mean and SD) across postdose time-points are shown in Figure 1. The geometric mean of the amisulpride peak plasma concentrations (Cmax) for all subjects after the first infusion was 411 ng/mL (95% CI, 352–480) in the absence of ondansetron and 409 ng/mL (95% CI, 341–492) in the presence of ondansetron; after the second amisulpride infusion, it was 475 ng/mL (95% CI, 409–553). Mean time of Cmax occurrence (Tmax) was 2–3 minutes after the start of infusion in all periods. Total amisulpride exposure (AUC[0]) after a single 10-mg dose of amisulpride, given concomitantly with ondansetron, was 289 ng·h/mL (SD, 51); after two 10-mg doses (in the absence of ondansetron), it was proportionally larger at 545 ng·h/mL (SD, 150), of which 136 ng·h/mL (25%) was attributable to the first 2 hours (ie, before the second dose). The mean clearance of amisulpride was 39 L/h (SD, 9) when given alone and 36 L/h (SD, 6) when given with ondansetron; the mean volume of distribution at steady state was 157 L (SD, 37) when given alone and 150 L (SD, 40) when given with ondansetron. The mean Cmax of ondansetron was 85 ng/mL (SD, 81), and AUC(0–) was 142 ng·h/mL (SD, 142).

Figure 1.:
Amisulpride plasma concentration profile when dosed alone and when dosed with ondansetron. Markers show arithmetic mean plasma concentration at each time-point, and error bars represent standard deviation around the mean.

Mean ΔHR on active treatment followed the pattern seen in the placebo period (Figure 2), and mean placebo-corrected change-from-baseline HR (ΔΔHR) was therefore within ±1.9 bpm at all postdose time-points, with the highest and lowest values both observed after dosing with amisulpride plus ondansetron.

Figure 2.:
ΔHR across treatment periods. ΔHR indicates change-from-baseline heart rate.

After dosing with amisulpride alone, the largest mean ΔQTcF reached 5.3 and 5.4 milliseconds 5 and 10 minutes after the first infusion and 5.7 and 6.1 milliseconds 5 and 20 minutes after the second infusion (125 and 140 minutes). After concomitant administration of amisulpride and ondansetron, the largest mean ΔQTcF was somewhat higher: 8.1 milliseconds 5 minutes after the start of the infusion (Figure 3A). Mean placebo-corrected ΔQTcF (ΔΔQTcF) values followed a similar pattern. After dosing with amisulpride only, mean ΔΔQTcF peaked at 5.2 milliseconds (90% CI, 3.53–6.96) 10 minutes after the first infusion and at 8.0 milliseconds (90% CI, 5.49–10.58) after the second infusion (130 minutes). After dosing with amisulpride plus ondansetron, mean ΔΔQTcF reached 7.3 milliseconds (90% CI, 5.48–9.16) at 5 minutes after the start of the first infusion (Figure 3B).

Figure 3.:
A, ΔQTcF. B, ΔΔQTcF. ΔQTcF indicates change-from-baseline QT interval corrected for heart rate using Fridericia’s formula; ΔΔQTcF, placebo-corrected change-from-baseline QT interval corrected for heart rate using Fridericia’s formula.

In the C-QTc analysis with amisulpride as analyte, a model including both the original term and a quadratic term for concentrations was fitted. The quadratic term was found to be statistically significant at the 0.05 level (P < .0001), thereby indicating that a linear model may not be appropriate. The slope of the C-ΔQTc relationship was shallow and statistically significant: 0.006 ms/ng/mL (90% CI, 0.002–0.010; P = .0149); with a statistically significant treatment effect–specific intercept of 2.6 milliseconds (90% CI, 2.1–3.2; P < .0001; Figure 4A, B and Table 1). The goodness-of-fit plot (Figure 4B) shows that the QT effect is somewhat underestimated at moderate plasma concentrations, whereas the observed QTc effect at high concentrations is captured. The effect on ΔΔQTcF can be predicted to 5.7 milliseconds (90% CI, 3.78–7.57) at the geometric mean Cmax in the amisulpride-only treatment period (501 ng/mL). Using this model, a predicted effect on ΔΔQTcF above 10 milliseconds can be excluded, as evidenced by the upper bound of the 2-sided 90% CI, for amisulpride plasma concentrations up to approximately 760 ng/mL (likely to be reached in <10% of subjects). Additional C-QTc models were also explored, with both ondansetron and amisulpride as analytes in the same model and on data from the amisulpride-alone treatment period. None of these models provided a better fit to the data, and predictions of the QTc effect (ΔΔQTcF) were similar.

Table 1. - Concentration-QTc Analysis of Amisulpride and Ondansetron and Associated ΔΔQTcF Prolongation
Parameter Estimate SE df t Value P Value 90% CI
Treatment effect (ms) 2.64 0.3446 1270.8 7.67 <.0001 2.075–3.210
Amisulpride slope (ms/ng/mL) 0.0059 0.0023 32.9 2.57 .0149 0.00203–0.00984
Centered baseline effect (ms) −0.23 0.0338 110.6 −6.71 <.0001 −0.283 to −0.171
1-min postdose effect (ms) 0.13 1.0716 33.4 0.12 .9028 −1.681 to 1.945
5-min postdose effect (ms) 1.87 1.0538 31.2 1.78 .0850 0.088–3.661
10-min postdose effect (ms) 1.96 1.0369 29.3 1.89 .0683 0.202–3.724
20-min postdose effect (ms) 1.46 1.0354 29.1 1.41 .1693 −0.300 to 3.218
30-min postdose effect (ms) 0.33 1.0374 29.4 0.31 .7554 −1.436 to 2.088
60-min postdose effect (ms) 0.024 1.0373 29.4 0.02 .9813 −1.7373 to 1.7862
120-min postdose effect (ms) −1.92 1.0382 29.5 −1.85 .0739 −3.687 to −0.161
121-min postdose effect (ms) −1.77 1.0478 30.6 −1.69 .1013 −3.548 to 0.007
125-min postdose effect (ms) 0.39 1.0386 29.5 0.38 .7087 −1.372 to 2.156
130-min postdose effect (ms) −1.56 1.0366 29.3 −1.51 .1429 −3.322 to 0.200
140-min postdose effect (ms) 0.16 1.0359 29.2 0.15 .8790 −1.601 to 1.919
150-min postdose effect (ms) −0.51 1.0364 29.3 −0.49 .6259 −2.271 to 1.250
180-min postdose effect (ms) 0.54 1.0372 29.3 0.52 .6062 −1.221 to 2.302
240-min postdose effect (ms) 0.12 1.0380 29.4 0.12 .9078 −1.642 to 1.884
360-min postdose effect (ms) −5.40 1.0389 29.5 −5.19 <.0001 −7.159 to −3.631
Abbreviations: ΔΔQTcF, placebo-corrected change-from-baseline QT interval corrected for heart rate using Fridericia’s formula; CI, confidence interval; df, degrees of freedom; QTc, QT interval corrected for heart rate; SE, standard error.

Figure 4.:
A, Scatter plot of observed amisulpride plasma concentrations and ΔΔQTcF by subject (amisulpride only model). The solid red line with dashed red lines denotes the model-predicted mean ΔΔQTcF with 90% CI. The blue squares and red triangles denote the pairs of observed amisulpride plasma concentrations and ΔΔQTcF by subjects for the amisulpride and amisulpride + ondansetron treatment periods, respectively. The individually placebo-adjusted ΔΔQTcFi,j equals the individual ΔQTcFi,j for subject i administered with amisulpride at time-point j minus the estimation of time at time-point j (ie, time effect). B, Model predicted and observed ΔΔQTcF (mean and 90% CI). The red-filled circles with vertical bars denote mean ΔΔQTcF with 90% CI displayed at the median plasma concentration within each decile for amisulpride. The solid black line with gray-shaded area denotes the model-predicted mean ΔΔQTcF with 90% CI. The horizontal red line with notches shows the range of amisulpride concentrations divided into deciles. The area between each decile represents the point at which 10% of the data are present; the first notch to second notch denotes the first 10% of the data, and the second notch to third notch denotes the 10%–20% of the data and so on. ΔQTcF indicates change-from-baseline QT interval corrected for heart rate using Fridericia’s formula; ΔΔQTcF, placebo-corrected change-from-baseline QT interval corrected for heart rate using Fridericia’s formula; CI, confidence interval.

Categorical analysis showed no clinically significant outliers. A QTcF between 450 and 480 milliseconds was recorded in 1 subject after dosing with amisulpride alone (at 7 time-points) and 1 after dosing with amisulpride plus ondansetron (at 2 time-points). No subject had a change-from-baseline QTcF >30 milliseconds. There were occasional observations of treatment-emergent changes of T-wave morphology across treatments, including flat T-waves in 1, 2, and 1 subjects after dosing with amisulpride, amisulpride plus ondansetron, and placebo, respectively.

Amisulpride at the studied doses and concentrations did not have an effect on cardiac conduction, that is, the PR and QRS intervals. Mean ΔPR was negative at all postdosing time-points and generally followed the pattern observed in placebo subjects. Mean ΔQRS values ranged from −0.3 to +0.7 milliseconds. Mean placebo-corrected change-from-baseline PR (ΔΔPR) and placebo-corrected change-from-baseline QRS (ΔΔQRS) values were within ±1.4 and ±0.7 milliseconds, respectively, across all postdosing time-points. There were no PR or QRS outliers.

Table 2. - TEAEs
Amisulpride 10 mg (2 Doses), N = 29 Amisulpride 10 mg + Ondansetron 4 mg, N = 30 Placebo, N = 30
Subjects, n (%) Events Subjects, n (%) Events Subjects, n (%) Events
All TEAEs 9 (31.0) 19 11 (36.7) 18 15 (50.0) 23
Mild TEAEsa 8 (27.6) 12 10 (33.3) 16 12 (40.0) 18
Moderate TEAEsa 5 (17.2) 7 2 (6.7) 2 5 (16.7) 5
Severe TEAEsa 0 (0.0) 0 0 (0.0) 0 0 (0.0) 0
Serious TEAEs 0 (0.0) 0 0 (0.0) 0 0 (0.0) 0
TEAEs leading to early termination 0 (0.0) 0 0 (0.0) 0 1 (3.3) 1
Deaths 0 (0.0) 0 0 (0.0) 0 0 (0.0) 0
Individual TEAEs
Allergy to arthropod bite 0 (0.0) 0 (0.0) 1 (3.3)
Back pain 1 (3.4) 1 (3.3) 0 (0.0)
Catheter site swelling 0 (0.0) 0 (0.0) 1 (3.3)
Constipation 0 (0.0) 1 (3.3) 0 (0.0)
Cough 0 (0.0) 0 (0.0) 1 (3.3)
Dizziness 0 (0.0) 2 (6.7) 1 (3.3)
Headache 4 (13.8) 3 (10.0) 2 (6.7)
Hordeolum (stye) 0 (0.0) 0 (0.0) 1 (3.3)
Infusion site discomfort 0 (0.0) 1 (3.3) 0 (0.0)
Infusion site pain 8 (27.6) 8 (26.7) 12 (40.0)
Nausea 1 (3.4) 0 (0.0) 0 (0.0)
Palpitations 0 (0.0) 0 (0.0) 1 (3.3)
Stomatitis 1 (3.4) 0 (0.0) 0 (0.0)
Abbreviation: TEAE, treatment-emergent adverse event.
aDefinitions used for assessment of severity were: mild—minimal discomfort and no interference with everyday activities; moderate—sufficient discomfort to interfere with normal everyday activities; and severe—prevents normal everyday activities and requires treatment or other intervention.

Details of treatment-emergent adverse events are provided in Table 2. There were no severe or life-threatening adverse events, serious adverse events, or deaths. The sole discontinuation occurred after administration of placebo. Most of the events (46/60, 77%) were judged mild by the investigator. The events judged moderate were infusion site pain, occurring in 4 subjects during the amisulpride-only period, 2 during the amisulpride plus ondansetron period, and 4 during the placebo period; and headache (2 subjects, amisulpride only) and hordeolum, or stye (1 subject, placebo).


This study demonstrated that a 10-mg dose of IV amisulpride has no effect on HR or EKG parameters, and that a clinically relevant effect on the QT interval can be excluded, even with concomitant ondansetron administration.

A second dose of amisulpride 2 hours after the first had essentially the same pharmacokinetics, with a slightly higher peak plasma concentration due to the residual levels from the first dose, and a similar QT effect. The decay of the QT effect (ΔΔQTcF; Figure 3B) appeared somewhat slower after the second dose, with a mean effect of around 5 milliseconds 60 minutes after the second infusion. However, the CI around the mean at this time-point was wide, overlapping with that at the corresponding time after the first infusion, so the small difference may be due to variability. When amisulpride was coadministered with ondansetron, a somewhat larger mean peak effect on ΔΔQTcF was observed (7.3 milliseconds), compared to amisulpride alone (5.2 milliseconds). This may result from somewhat higher end-of-infusion amisulpride plasma concentrations (Figure 1), but a small additive QT effect of ondansetron cannot be excluded.

The data are generally consistent with those published previously on both amisulpride and ondansetron. For example, the maximal QTc effect of 5-mg amisulpride was 5.0 milliseconds (90% CI, 2.8–7.1) in a previous TQT study.10 Based on a linear, mixed-effects concentration-effect model derived from data on both 5- and 40-mg doses, the predicted QT effect of 10 mg was around 9 milliseconds. Similarly, pharmacokinetic data in the ondansetron prescribing information17 (Cmax of 102 ng/mL, AUC[0–∞] of 156 ng·h/mL) closely accord with our findings. Also consistent is the absence of outliers, with no absolute QTcF values above 480 milliseconds nor any changes from baseline >40 milliseconds, even at a 40-mg dose.

Our data contrast interestingly with those published on ondansetron combined with droperidol.18 In 16 healthy volunteers given 4-mg ondansetron over 2 minutes and 1-mg droperidol over 10 seconds, the mean maximal ΔΔQTcF was 25 milliseconds after droperidol, 17 milliseconds after ondansetron, and 28 milliseconds after the combination. It is unclear why the QT effect of ondansetron was so high in that study, although interestingly the median Cmax recorded for ondansetron was 175 ng/mL, which is more than twice that observed in our study. Given that difference, caution is warranted in predicting the possible effect of higher peak plasma ondansetron concentrations on any QT-related interaction with amisulpride. However, the small additive effect of ondansetron above that of droperidol (3 milliseconds) matches our findings.

In view of the possibility of amisulpride and ondansetron coadministration in PONV management, the data from this study may assist prescribers. Most patients in the United States given PONV prophylaxis receive 4-mg ondansetron.19 Ondansetron is known to have a dose-dependent QT effect, the largest ΔΔQTcF values being 5.8 and 19.6 milliseconds for 15-minute infusions of 8 and 32 mg, respectively.15 A randomized trial showed that a single 5-mg amisulpride dose is effective as PONV prophylaxis when combined with another antiemetic, such as ondansetron or dexamethasone,5 and concomitant use of amisulpride and ondansetron can, therefore, be expected in practice. Furthermore, amisulpride at 10 mg (but not 5 mg) has demonstrated effectiveness in treating PONV in patients who have failed ondansetron prophylaxis.7 Currently, ondansetron redosing is common in that setting,19 even though it has been shown to be ineffective20 and is not recommended in the ondansetron label.17 It is, therefore, conceivable that some clinicians could continue redosing ondansetron and add in amisulpride when rescuing such patients. The data from this study provide some reassurance that such combination use would not represent a significant QT risk.

The study followed published guidance on the evaluation of QT effects and, in particular, the recommendations in a white paper on concentration-QT modeling published by a specialist group of FDA and industry experts.12 Even though predictions of the effect of amisulpride on ΔΔQTcF using several explored models were consistent with the observed by-time-point effect, none of the models provided a good fit to the data and all had relatively large intercept, which is an indicator of a misspecified model. This is probably due to the relatively small QT effect, in combination with an observed, albeit short, delay between Cmax and the peak QT effect. The test for linearity indicates that a nonlinear model may improve the model fit, but the scatter plot (Figure 4A) and the goodness-of-fit plot (Figure 4B) clearly show that nonlinear models, for example, maximum effect of drug (Emax) models, are likely to underestimate the QT effect at high amisulpride plasma concentrations. Given the small predicted QTc effect with a linear model and to stay on the conservative side, we have therefore, in line with regulatory practice,12 not further explored nonlinear models. It is, however, acknowledged that the linear model seems to somewhat underestimate the effect within the amisulpride concentration range from ≈100 to ≈250 ng/mL. Predictions of the QT effect at plasma concentrations above 500 ng/mL based on the results of the C-QTc analysis in this study should, however, only be done with caution.

One limitation of this study is that the subjects were volunteers rather than surgical patients undergoing a range of procedures and therapies which could prolong QT. These may constitute a risk of dysrhythmia, especially in patients with preexisting vulnerability, and caution is always warranted in the perioperative setting. While it is possible that amisulpride may have a different effect in anesthetized patients, it appears from this study that amisulpride does not per se contribute a meaningful risk in respect of QT prolongation.

Another potential limitation is the absence of a positive control, such as moxifloxacin. However, the high precision of the methods used for EKG extraction and QT analysis, the volume of data points at relevant concentrations, and the concordance of the data with the previous TQT study provide confidence regarding assay sensitivity.

In conclusion, unlike several other antiemetics, including the dopamine antagonists droperidol, haloperidol and metoclopramide, and the 5-hydroxytryptamine (serotonin), type 3 (5HT3)-antagonist dolasetron, amisulpride does not appear to have a clinically meaningful effect on QT when used at doses shown to be effective for preventing and treating PONV, even when coadministered with ondansetron. A second amisulpride dose after 2 hours has essentially the same pharmacokinetics and QT profile as the first.


Name: Gabriel M. Fox, MB BChir.

Contribution: This author was study director and helped write the manuscript.

Conflicts of Interest: G. M. Fox is an employee and stockholder of Acacia Pharma Ltd.

Name: Muna Albayaty, MBChB, FFPM, MSc.

Contribution: This author was study principal investigator and helped review the manuscript.

Conflicts of Interest: None.

Name: Joanna L. Walker, BSc.

Contribution: This author was study manager and helped review the manuscript.

Conflicts of Interest: J. L. Walker is an employee and stockholder of Acacia Pharma Ltd.

Name: Hongqi Xue, PhD.

Contribution: This author was study statistician and helped write the manuscript.

Conflicts of Interest: None.

Name: Borje Darpo, MD, PhD.

Contribution: This author advised on electrocardiogram (EKG) methods and helped write the manuscript.

Conflicts of Interest: B. Darpo owns stock and is eligible for stock options in ERT.

This manuscript was handled by: Ken B. Johnson, MD.


1. Gan TJ, Diemunsch P, Habib AS, et al.; Society for Ambulatory Anesthesia. Consensus guidelines for the management of postoperative nausea and vomiting. Anesth Analg. 2014;118:85–113.
2. Habib AS, Gan TJ. The use of droperidol before and after the Food and Drug Administration black box warning: a survey of the members of the Society of Ambulatory Anesthesia. J Clin Anesth. 2008;20:35–39.
3. Kranke P, Eberhart L, Motsch J, et al. I.V. APD421 (amisulpride) prevents postoperative nausea and vomiting: a randomized, double-blind, placebo-controlled, multicentre trial. Br J Anaesth. 2013;111:938–945.
4. Gan TJ, Kranke P, Minkowitz HS, et al. Intravenous amisulpride for the prevention of postoperative nausea and vomiting: two concurrent, randomized, double-blind, placebo-controlled trials. Anesthesiology. 2017;126:268–275.
5. Kranke P, Bergese SD, Minkowitz HS, et al. Amisulpride prevents postoperative nausea and vomiting in patients at high risk: a randomized, double-blind, placebo-controlled trial. Anesthesiology. 2018;128:1099–1106.
6. Candiotti KA, Kranke P, Bergese SD, et al. Randomized, double-blind, placebo-controlled study of intravenous amisulpride as treatment of established postoperative nausea and vomiting in patients who have had no prior prophylaxis. Anesth Analg. 2019;128:1098–1105.
7. Habib AS, Kranke P, Bergese SD, et al. Amisulpride for the rescue treatment of postoperative nausea or vomiting in patients failing prophylaxis: a randomized, placebo-controlled phase III Trial. Anesthesiology. 2019;130:203–212.
8. International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) Topic E14. The Clinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential for Non-Antiarrhythmic Drugs. 2005. Available at: Accessed November 5, 2019.
9. Darpo B. The thorough QT/QTc study 4 years after the implementation of the ICH E14 guidance. Br J Pharmacol. 2010;159:49–57.
10. Täubel J, Ferber G, Fox G, Fernandes S, Lorch U, Camm AJ. Thorough QT study of the effect of intravenous amisulpride on QTc interval in Caucasian and Japanese healthy subjects. Br J Clin Pharmacol. 2017;83:339–348.
11. Darpo B, Fossa AA, Couderc JP, et al. Improving the precision of QT measurements. Cardiol J. 2011;18:401–410.
12. Garnett C, Bonate PL, Dang Q, et al. Scientific white paper on concentration-QTc modeling. J Pharmacokinet Pharmacodyn. 2018;45:383–397.
13. International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) Topic E14. The Clinical Evaluation of QT/QTc Interval Prolongation and Proarrhythmic Potential for Non-Antiarrhythmic Drugs, Questions & Answers (R3). 2015. Available at: Accessed November 5, 2019.
14. Denney WE, Duvvuri S, Buckeridge C. Simple, automatic noncompartmental analysis: The PKNCA R Package. J Pharmacokinet Pharmacodyn. 2015;42:S65.
15. Zuo P, Haberer LJ, Fang L, Hunt TL, Ridgway D, Russo MW. Integration of modeling and simulation to support changes to ondansetron dosing following a randomized, double-blind, placebo-, and active-controlled thorough QT study. J Clin Pharmacol. 2014;54:1221–1229.
16. Darpo B, Garnett C, Keirns J, Stockbridge N. Implications of the IQ-CSRC Prospective Study: time to revise ICH E14. Drug Saf. 2015;38:773–780.
17. FDA. Full Prescribing Information, Zofran (ondansetron hydrochloride) Injection for Intravenous use. 2017. Available at: Accessed November 5, 2019.
18. Charbit B, Alvarez JC, Dasque E, Abe E, Démolis JL, Funck-Brentano C. Droperidol and ondansetron-induced QT interval prolongation: a clinical drug interaction study. Anesthesiology. 2008;109:206–212.
19. Macario A, Claybon L, Pergolizzi JV. Anesthesiologists’ practice patterns for treatment of postoperative nausea and vomiting in the ambulatory Post Anesthesia Care Unit. BMC Anesthesiol. 2006;6:6.
20. Kovac AL, O’Connor TA, Pearman MH, et al. Efficacy of repeat intravenous dosing of ondansetron in controlling postoperative nausea and vomiting: a randomized, double-blind, placebo-controlled multicenter trial. J Clin Anesth. 1999;11:453–459.
Copyright © 2020 International Anesthesia Research Society